microsoft/onnxruntime

Pull Requests Commits

Clarify WebGPU EP device selection

edgchen1 committed 3 days ago

875192ac

only use first EP device for WebGPU EP.

edgchen1 committed 3 days ago

cc0df426

Add WebGPU EP device diagnostics

edgchen1 committed 3 days ago

5f54d1e5

Add more paths for cuda plugin ep (#31688)

tianleiwu committed 3 days ago

Verified a86d86ad

onnxruntime-web JSEP deprecation announcement for developers (#31683)

edgchen1 committed 3 days ago

Verified 407638ad

Add onnxruntime/test/providers/webgpu to plugin-ep-webgpu/paths.txt (#31685)

edgchen1 committed 3 days ago

Verified 2d1cfa02

[WebGPU] Support odd-N subgroup matrix MatMul weights (#29893)

jchen10 committed 3 days ago

Verified 1fe713be

Update ort release skill for cuda plugin (#31669)

tianleiwu committed 3 days ago

Verified 7ad118ca

[CUDA] GroupQueryAttention: dequantize the K and V caches in one launch (#31480)

tianleiwu committed 3 days ago

Verified c003602a

[WebGPU] Accumulate MatMulNBits wide-tile in output_element_t (#29611)

daijh committed 3 days ago

Verified baa0d32b

[CUDA] Speed up the NVFP4 QMoE decode GEMV and enable it for MTP verify (#31159)

tianleiwu committed 3 days ago

Verified e56b1073

[CPU] Tighten cache_indirection shape contract in MultiHeadAttention (#31634)

nenad1002 committed 4 days ago

Verified fbaee34d

Guard CUDA LayerNorm/RMSNorm int32 offset range (#31650)

apsonawane committed 4 days ago

Verified 24f1aedc

[CUDA] Add build option for TensorRT fused attention cubins (#31632)

tianleiwu committed 4 days ago

Verified 25cc21c2

Add ort-release-notes skill to draft release notes (#31614)

edgchen1 committed 4 days ago

Verified 1648b645

[CUDA] PagedAttention: quantized KV cache, XQA decode, MLA, QK-Norm and head sink (#29912)

tianleiwu committed 4 days ago

Verified e644fd56

[CUDA] Add win-arm64 packaging and size options to plugin pipeline (#31635)

tianleiwu committed 4 days ago

Verified 6df2a551

[WebGPU plugin EP] Update release and packaging-related docs (#29059)

edgchen1 committed 4 days ago

Verified c0b1d6bd

Handle empty tensors in CUDA InstanceNormalization (#31647)

apsonawane committed 4 days ago

Verified 4bae5766

[DML EP] Fix wide string handling in OpKernelInfoWrapper::GetWideName (#31656)

miaobin committed 4 days ago

Verified 4e48876e

web: fix webpack/Terser release build crash in Web CI Pipeline (#31652)

Copilot committed 4 days ago

Verified b996de6b

[CUDA] MatMulBlockQuantizedFp8Weight: fold the W8A8 activation QDQ into the decode GEMV (#31481)

tianleiwu committed 5 days ago

Verified 064861b0

[onnxruntime-web] Plans to deprecate and remove WebGL and JSEP (#29716)

edgchen1 committed 5 days ago

Verified fef99e8d

[WebNN EP] Reuse shared WASM loader for Blob-backed external data (#31151)

Honry committed 5 days ago

Verified 30c3dc47

Throw error on negative Split tensor axis [CPU] (#31149)

nenad1002 committed 5 days ago

Verified dfa30b05

[Build] update build flags of cuda plugin CI to avoid OOM (#31622)

tianleiwu committed 5 days ago

Verified 2ed68162

[WebGPU EP] Support int64 for Tile and Concat (#31049)

miaobin committed 5 days ago

Verified f6d5554e

[Build] Fix cuda plugin linux build errors (#31616)

tianleiwu committed 5 days ago

Verified 5ea5e692

[Build] Fix Windows ARM64 CUDA plugin build (#31617)

tianleiwu committed 5 days ago

Verified edb8580d

[Build] Disable SM120 TMA kernels with MSVC and CUDA 13 (#31615)

tianleiwu committed 5 days ago

Verified 4cf0af68

Older